JP2011000184A - Magnetic resonance imaging apparatus and magnetic resonance imaging method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a magnetic resonance imaging apparatus imaging a cross section of a speech organ, which vibrates at such a high speed as a frequency of the sound in an audible range, as a moving image by using nuclear magnetic resonance.SOLUTION: The MRI apparatus 1000 includes a data storing part 258 for storing a voice signal corresponding to the periodic kinetic phase of the speech organ of a subject on a one-on-one level, and an image reconstruction part 260 for applying a vibration magnetic field to an RF coil 104 and generating tomographic moving images as frames in time series on receiving a detection signal. The image reconstruction part 260 acquires data for reconstructing tomographic images a plurality of times for each line along a lead-out axis in a k-space asynchronously with the kinetic motion of the speech organ, and selects the line corresponding to the same phase of the kinetic motion of the speech organ based on the voice signal at the timing for the respective frames of the moving images to reconstruct the cross-sectional image measured a plurality of times, so that the image of the phase is reconstructed.

Description

  The present invention relates to a configuration of a magnetic resonance imaging (MRI) apparatus for performing tomographic imaging of a living body and a magnetic resonance imaging method, and more particularly, to a tomographic object to be measured at a high frame rate. The present invention relates to a magnetic resonance imaging apparatus and a magnetic resonance imaging method capable of capturing moving images.

  Magnetic resonance imaging using the nuclear magnetic resonance phenomenon of the living body's brain and whole body cross-sections (tomograms) with respect to atoms in the living body, especially hydrogen nuclei, is used for human clinical imaging diagnosis, etc. Has been.

  When applying it to the human body, the magnetic resonance imaging has the following features, for example, compared to “X-ray CT” which is a similar tomographic image of the human body.

  (1) An image having a density corresponding to the distribution of hydrogen atoms and the signal relaxation time can be obtained. For this reason, the lightness and darkness corresponding to the difference in tissue properties is exhibited, and the difference in tissue is easily observed.

  (2) The bone has a lower density of hydrogen atoms than other parts. For this reason, it is easy to observe a part (inside the skull, spinal cord, etc.) surrounded by bones.

(3) Since it is not harmful to the human body like X-rays, it can be used in a wide range.
Such magnetic resonance imaging utilizes the magnetic properties of hydrogen nuclei (protons) that are most abundant in each cell of the human body and have the greatest magnetism. The movement in the magnetic field of the spin angular momentum responsible for the magnetism of the hydrogen nucleus is classically compared to the precession of the coma.

The following is a simple summary of the principles of nuclear magnetic resonance in this intuitive classic model.
The direction of the spin angular momentum of the hydrogen nuclei (the direction of the rotation axis of the coma) as described above is directed in a random direction in an environment without a magnetic field, but is directed in the direction of the lines of magnetic force when a static magnetic field is applied.

  When an electromagnetic wave (RF (Radio Frequency) pulse) is further superimposed in this state, the frequency of the electromagnetic field is a resonance frequency f0 = γB0 / 2π (γ: a coefficient specific to the substance) determined by the strength of the static magnetic field. The energy moves to the nucleus side by resonance, and the direction of the magnetization vector changes (precession increases). When the electromagnetic wave is cut in this state, the precession returns to the direction in the static magnetic field while returning the flip angle. By detecting this process from the outside with an antenna coil, an NMR signal can be obtained.

  Such a resonance frequency f0 is 42.6 × B0 (MHz) for hydrogen atoms when the strength of the static magnetic field is B0 (T).

  And MRI using such nuclear magnetic resonance is also used for observing the state of the articulatory organ of a human being uttered, as will be described below.

  Here, various methods have been used so far to visualize the dynamics of speech organs that perform dynamic movements such as the tongue and vocal cords.

  For articulatory motion, moving image imaging using ultrasonic waves, X-ray microbeams, and MRI (hereinafter referred to as “MRI movie”) (see Non-Patent Document 1, Non-Patent Document 2, and Non-Patent Document 3) and EMMA ( By using ElectroMagnetic Midsagittal Articulometer), it has become possible to understand the movement of the tongue and larynx that cannot be seen from outside the body.

  On the other hand, speech organs that vibrate at a very high speed, such as vocal cord vibrations and gum shaking sounds, could not be observed so far by the above method, limited by speech synchronization and size of the vibration part. For this reason, the vocal cords that vibrate at 100 Hz or higher have been observed from above using a laryngeal stroboscopic copy or a high-speed camera, or by flowing an artificial air flow through the extracted larynx (Non-Patent Document 4). Non-patent document 5). Endoscopic observation is mainly limited to observation from above, so that the lower lip of the vocal cords has a time interval that cannot be observed due to upper lip closure. The extracted larynx is difficult to perform in the case of a human body, and therefore it is difficult to reproduce the vocal cord movement that characterizes human speech / singing voice.

  A feature of MRI is that the cross-sectional shape of the uttering organ can be observed, and imaging in a plurality of cross-sections is also possible. Furthermore, there is an advantage that there is no concern about invasiveness. Taking advantage of these advantages, the MRI movie imaging method developed further has been used to observe articulatory motion. As a result, a moving image of 120 frames per utterance was obtained, and the articulatory organ was successfully visualized.

  However, imaging of organs that change faster than articulatory movement is difficult with the MRI movie imaging methods proposed so far. In the method of creating a moving image with a single utterance using the spiral k space (see Non-Patent Document 6), there is a problem that the frame rate is insufficient for vocal cord vibration or the like.

  In addition, the method in which the subject synchronizes the utterance during MRI movie imaging (see Non-Patent Document 1) cannot be applied as it is because the utterer cannot control closing / opening by vocal cord vibration or the like. There has also been proposed a method that does not require utterance synchronization by recording utterance voice of a subject at the time of imaging and rearranging data in k-space based on the utterance voice (see Non-Patent Document 2). It was difficult to animate a speech organ that vibrates at a frequency.

S. Masaki, MKTiede, K. Honda, Y. Shimada, I. Fujimoto, Y. Nakamura and N. Ninomiya: “MRI-based speech pro-duction study using a synchronized sampling method”, JA coust. Soc. Jpn. (E), 20, 5, pp. 375-379 (1999). M.Mohammad, E.Moore, J.N.Carter, C.H.Shadle and S. J. Gunn: “Using MRI to image the moving vocal tract during speech”, Proc. Eurospeech’97, Vol. 4, pp. 2027-2030 (1997). S. Narayanan, K. Nayak, S. Lee, A. Sethy and D. Byrd: “An approach to real-time magnetic resonance imaging for speech production”, J. Acoust. Soc. Am., 115, 4, pp. 1771-1776 (2004). T. Baer: “Investigation of phonation using excised larynxes”, PhD thesis, Massachusetts Institute of Technology (1975). D.A.Berry, D.W.Montequin and N.Tayama: “High-speed digital imaging of the medial surface of the vocal folds”, J. Acoust. Soc. Am., 110, 5, pp. 2539-2547 (2001). S. Narayanan, K. Nayak, S. Lee, A. Sethy and D. Byrd: “An approach to real-time magnetic resonance imaging for speech production”, J. Acoust. Soc. Am., 115, 4, pp. 1771-1776 (2004).

  Hereinafter, the operation and problems of the conventional MRI movie imaging method will be briefly described.

FIG. 13 is a conceptual diagram showing the operation principle of MRI.
In MRI, a subject 10 is laid in a static magnetic field coil for generating a static magnetic field, and information on the position of an observation cross section (slice) in the subject 10 is given to an observation signal by a magnetic field generated by a gradient magnetic field coil. This is called “slice selection”. Further, a gradient magnetic field for phase encoding that adds position information in the slice is applied at a predetermined timing. Further, an electromagnetic wave (RF pulse) is applied to the nucleus to be observed by the RF coil, and a reading gradient magnetic field for performing frequency encoding for specifying the position in the direction orthogonal to the one direction in the slice is provided. Applied.

  In this way, the external magnetic field, the RF wave, and the gradient magnetic field are given according to a predetermined sequence, so that a signal from the observation target hydrogen nucleus is received by the receiving coil, and data for one lead-out line in the k space is obtained. To be acquired. In general, data of all points in k-space is acquired by repeating such a sequence a predetermined number of times while changing the strength of the gradient magnetic field for phase encoding for each lead-out line.

  An image is reconstructed by performing fast Fourier transform (FFT) on the data in the k-space.

FIG. 14 is a diagram showing a trajectory of acquired data observed in the k space.
In FIG. 14, the direction of the vertical axis represents a change in phase due to phase encoding, and the interval of the vertical axis corresponds to the repetition time interval TR of RF pulse application. The horizontal axis indicates the frequency, and the interval corresponds to the readout sampling time Ts. In this figure, a line of data in the horizontal axis direction is referred to as a “lead outline”.

FIG. 15 is a conceptual diagram illustrating a configuration of a technique in which the subject performs the utterance synchronization described above.
The subject 10 repeats the same utterance task according to the notification sound of the utterance timing. On the MRI apparatus side, data acquisition is performed in accordance with the trigger signal output from the external trigger generation unit in accordance with the timing from the notification sound generation unit of the utterance timing.

  Here, in one data acquisition period, the gradient magnetic field for phase encoding is a constant value, and data is acquired in the readout direction at different repetition timings. In the example of FIG. 15, the case where such data acquisition is repeated 128 times is shown.

  In the 128-time data acquisition, when data in the readout direction at the same timing from the external trigger are combined, data corresponding to different phase encoding sizes can be obtained from the external trigger at the same timing, and one imaging field is obtained. K-space data corresponding to the FOV is acquired. In this manner, k-space data corresponding to the imaging field FOV at the timing of each elapsed time is collected for each imaging field with the external trigger as a reference. When the k-space data collected in this way is subjected to fast Fourier transform, image data for each imaging field is reconstructed.

  Since such image data is obtained at the timing of each elapsed time with the external trigger as a reference, this constitutes each frame of the moving image.

  However, as described above, with such a technique, it is difficult to image a moving organ as a moving image faster than articulatory movement because the speaker itself cannot control closing and opening with vocal cord vibration or the like. There was a problem.

  For example, imaging of cross-sections of these organs as a moving image was difficult in gum ringing sound (so-called “rolling tongue”. Trrrr ... sound), vocal cord vibration, lip vibration when wearing a mouthpiece, and the like.

  In other words, more generally, it has been difficult to image a cross section of a vibrator that vibrates at a high speed about the fundamental frequency of sound as a moving image.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to use a nuclear magnetic resonance to vibrate at a high speed, for example, at the fundamental frequency of audio in the audible region. It is an object to provide a magnetic resonance imaging apparatus and a magnetic resonance imaging method capable of imaging a cross-section of the above as a moving image.

  According to one aspect of the present invention, there is provided a magnetic resonance imaging apparatus for detecting a detection signal caused by nuclear magnetic resonance from a transducer to be measured and generating a tomographic motion image of the measurement target. , A static magnetic field applying means for applying a static magnetic field to the measurement target, and a magnetic field modulated in such a manner that the detection signal has the position information of the nucleus that emits the detection signal within the selected cross section of the measurement target Gradient magnetic field applying means for applying to the object, electromagnetic field transmitting / receiving means for applying an electromagnetic field to the object to be measured and detecting a detection signal from the object to be measured, and the periodicity of the vibrator of the object to be measured Storage means for storing an output signal from the transducer corresponding to the phase of motion one-to-one, and a tomography for applying an electromagnetic field to the electromagnetic field transmitting / receiving means and receiving a detection signal to generate a tomographic moving image as a time-series frame Shooting control A tomography control means, wherein the tomography control means obtains data for reconstructing the cross-sectional image multiple times for each row along the k-space readout axis, asynchronously with the motion of the transducer; By selecting the row corresponding to the same phase of the motion of the transducer based on the output signal at the timing of each frame of the moving image that reconstructs the cross-sectional image being measured multiple times, the image of the phase Image reconstructing means for reconstructing the image.

  Preferably, the apparatus further comprises timing detection means for acquiring information for specifying the measurement timing for each row along the lead-out axis, and the storage means reconstructs the cross-sectional image together with information indicating the measurement timing. To store data.

Preferably, the output signal is an audio signal from the vibrator.
Preferably, a band pass filter is further provided for separating the audio signal from noise from the magnetic resonance imaging apparatus.

  According to another aspect of the present invention, there is provided a magnetic resonance imaging method for detecting a detection signal caused by nuclear magnetic resonance from a transducer to be measured and generating a tomographic motion image of the measurement target. Applying a static magnetic field to the object to be measured, applying a magnetic field modulated so that the detection signal has the position information of the nucleus emitting the detection signal in the selected cross section of the object to be measured; A step of detecting a detection signal from the object to be measured by applying an electromagnetic field to the object to be measured, and an output from the vibrator having a one-to-one correspondence with the phase of the periodic movement of the vibrator to be measured. A step of storing a signal, and a step of generating a tomographic motion image as a time-series frame by applying an electromagnetic field to an object to be measured, receiving a detection signal, Is asynchronous, The step of acquiring data for reconstructing the plane image a plurality of times for each row along the k-space readout axis, and the timing of each frame of the moving image for reconstructing the cross-sectional image being measured a plurality of times, Selecting a row corresponding to the same phase of motion of the transducer based on the output signal to reconstruct an image of that phase.

  According to the present invention, for example, a cross section of a vibrator that vibrates at a high speed about the fundamental frequency of sound can be imaged as a moving image.

It is a functional block diagram which shows the structure of the MRI apparatus 1000 of an example of the magnetic resonance imaging apparatus which concerns on this invention. 3 is a conceptual diagram for explaining image reconstruction processing by the MRI apparatus 1000. FIG. It is a conceptual diagram for demonstrating an image reconstruction process compared with the conventional image reconstruction process. When a vocal cord is considered as a vocal organ, it is a figure which shows the time change of the cross-sectional shape of a vocal cord, a speech waveform, and a signal waveform after passing a speech signal through a band transmission filter. It is a figure which shows the pulse sequence of the MRI apparatus 1000. FIG. 4 is a flowchart for explaining the flow of processing of the MRI apparatus 1000. It is a figure explaining the components which comprise a phantom. It is a figure which shows the external appearance and sectional drawing of a phantom. It is a figure which shows the spectrogram of the noise which arises from the MRI apparatus 1000, and the sound which arises from a vibrator | oscillator. It is a figure which shows the recorded gradient magnetic field and an audio | voice. It is a figure which contrasts and shows the image of a MRI movie and a high-speed camera. It is a figure which shows the data for 26-50 frames among 50 frames. It is a conceptual diagram which shows the principle of operation of MRI. It is a figure which shows the trajectory of the income data observed in k space. It is a conceptual diagram which shows the structure of the method of performing utterance synchronization.

Embodiments of the present invention will be described below with reference to the drawings.
[Configuration and operation of magnetic resonance imaging apparatus]
FIG. 1 is a functional block diagram showing a configuration of an MRI apparatus 1000 as an example of a magnetic resonance imaging apparatus according to the present invention.

  As described below, the MRI apparatus 1000 can increase the time resolution based on the MRI movie technology.

  The MRI apparatus 1000 shown in FIG. 1 includes a platform 12 for supporting the subject 10, a static magnetic field coil 100 for generating a static magnetic field, and the position of an observation cross section (slice) in the subject as will be described later. Gradient magnetic field coil 102 for giving information on the position in the slice to the observation signal, RF coil 104 for outputting an electromagnetic wave (RF pulse) to the nucleus to be observed, and a signal from the nucleus to be observed are received And a tomography control unit 200 for controlling the coils 100 to 104 and generating a tomographic image based on a signal received by the receiving coil 106.

  Although the RF coil 104 and the receiving coil 106 are described as being separate coils, the same coil can be shared for transmission and reception.

  An optical microphone 108 for acquiring the speech voice of the subject is installed in the vicinity of the subject 10, and a solenoid type coil 110 for detecting a change in the gradient magnetic field is provided in the gantry of the MRI apparatus. Is provided. The reason for using the optical microphone is that it is possible to capture sound even in a strong magnetic field environment because no metal is used.

Further, the tomography control unit 200 is configured to input from the RF coil 104, for example, every measurement for each subject 10 based on the input unit 210 for inputting an instruction from the user and the instruction from the input unit 210. The adjustment value of the intensity of the electromagnetic field to be applied and the signal intensity detected by the receiving coil 106 and the result (calibration value) of the adjustment of the value f ₀ of the measured resonance frequency (hereinafter referred to as the center frequency) are stored as tuning values. A tuning value storage unit 220. Here, the reason why such tuning is performed for each subject 10 is that the subject 10 needs to be adjusted because the environment of the magnetic field in the coil slightly changes.

  The tomography control unit 200 further includes a control unit 230 for controlling the tuning operation and measurement operation as described above, and an RF pulse for applying an RF pulse to the RF coil 104 under the control of the control unit 230. A transmission unit 240, an MRI signal amplification unit 250 for amplifying a signal from the reception coil 106 to obtain a detection signal, an audio signal amplification unit 252 for receiving and amplifying the audio signal from the optical microphone 108, A data storage unit 258 that stores the measurement data corresponding to the detection signal from the MRI signal amplification unit 250 and the audio signal data from the audio signal amplification unit 252 together with the data acquisition timing under the control of the control unit 230, as will be described later. , For obtaining only a signal of a predetermined frequency band among the audio signals stored in the data storage unit 258 In order to reconstruct the cross-sectional image data of the cross-section to be observed by performing Fourier transform processing based on the band-pass filter 254, the measurement data from the data storage unit 258, and the audio signal data that has passed through the band-pass filter 254. The image reconstruction unit 260 and a display unit 270 for displaying the cross-sectional image reconstructed based on the information from the image reconstruction unit 260 are provided. The cross-sectional image data reconstructed in this way is stored in the data storage unit 258.

  Here, preferably, in the image reconstruction process in the image reconstruction unit 260, the information such as the calibration value of the center frequency stored in the tuning value storage unit 220 is used to improve the image quality. Is possible. In FIG. 1, the change in the gradient magnetic field is detected by the solenoid coil 110, but the control unit 230 may acquire the timing at which the gradient magnetic field changes by a control signal for controlling the gradient magnetic field. Good.

  Although not particularly limited, the control unit 230 and the image reconstruction unit 260 are realized by the same arithmetic processor that operates based on software stored in a storage device (not shown) or individual arithmetic processors corresponding to the respective processes. It is possible.

  Here, more specifically, the static magnetic field coil 100 is composed of, for example, four air-core coils, and a combination thereof creates a uniform magnetic field to give orientation to the spins of hydrogen nuclei in the body of the subject 10. .

  The RF coil 104 emits a high frequency to excite nuclei in the body of the subject 10, and the receiving coil 106 detects a detection signal (echo signal) caused by the generated nuclear magnetic resonance.

  The gradient coil 102 includes three sets of gradient coils of X, Y, and Z (not shown), and the Z coil generates a slice selective gradient magnetic field for limiting the resonance surface by tilting the magnetic field intensity in the Z direction when excited. The Y coil adds a short time gradient immediately after the application of the magnetic field in the Z direction to apply phase modulation proportional to the Y coordinate to the detection signal (phase encoding), and the X coil subsequently adds a gradient during data collection. Thus, frequency modulation proportional to the X coordinate is given to the detection signal (frequency encoding).

  That is, when a high frequency electromagnetic field having a resonance frequency is applied to the subject 10 in a state in which a Z-axis gradient magnetic field is added to the static magnetic field through the RF coil 104, the hydrogen nuclei in the portion where the strength of the magnetic field is in the resonance condition , Selectively excited to begin to resonate. Hydrogen nuclei in a portion that matches the resonance condition (for example, a tomography of a predetermined thickness of the subject 10) are excited, and spins rotate together. When the excitation pulse is stopped, an electromagnetic wave radiated from the rotating spin is induced in the receiving coil 106 this time, and this signal is detected for a while. By this signal, a tissue containing hydrogen atoms in the body of the subject 10 is observed. And in order to know the transmission position of a signal, it is the structure of adding a gradient magnetic field of X and Y, and detecting a signal.

  The control unit 230 measures the detection signal while repeatedly applying the excitation signal, and the image reconstruction unit 260 reduces the resonance frequency to the X coordinate by Fourier transform calculation, restores the Y coordinate, obtains an image, An image corresponding to the display unit 270 is displayed.

  FIG. 2 is a conceptual diagram for explaining image reconstruction processing by the MRI apparatus 1000 shown in FIG.

  Referring to FIG. 2, subject 10 utters a predetermined utterance task. At this time, the data acquisition timing and the utterance timing are not synchronized. On the MRI apparatus 1000 side, the speech sound acquired by the optical microphone 108 is stored in the data storage unit 258 together with the acquisition timing information, and the control unit 230 receives the data acquisition timing from the gradient magnetic field detected by the solenoid coil 110. The MRI signal is stored in the storage unit 258 together with the acquisition timing information.

  Here, the “acquisition timing information” may be time information itself, and as described later, “speech speech” and “MRI signal” are recorded on the same time axis. It is good also as recording simultaneously on a different channel.

  The acquisition of the MRI signal is not particularly limited. For example, according to a pulse sequence of a so-called “gradient echo method”, the magnitude of the gradient magnetic field for phase encoding is sequentially changed in the k-space corresponding to one imaging field. Get the data. When data acquisition in the k space corresponding to one imaging field is completed, such data acquisition is repeated a predetermined number of times.

  Thereafter, based on the acquired audio data, k-space data about the frame timing of each moving image is obtained by sorting and combining the k-space data in which the phases of the vibrations of the audio match at the timing of each frame constituting the moving image. Get the data. Image data of each frame of a moving image can be reconstructed by performing fast Fourier transform on the k-space data obtained in this way for each imaging field.

  FIG. 3 is a conceptual diagram for explaining the image reconstruction process in comparison with the conventional image reconstruction process.

  Referring to FIG. 3, the image reconstruction process in the MRI apparatus 1000 is to continuously scan a target transducer, collect a lead-out line in k-space at each vibration timing, and image it. As a result, the MRI movie can be imaged even when the uttered organ that vibrates whose phase cannot be controlled by the utterer himself / herself is targeted.

  FIG. 3 shows a comparison between a general MRI still image imaging method (a) and a method (b) of the MRI apparatus 1000.

Specifically, the imaging method of the MRI apparatus 1000 is as follows.
1) Simultaneously measure and store the imaging timing and the position of the transducer at that time. The imaging is continuous scanning, and the speaker vibrates the speaking organ at a constant frequency at his own timing. As a method for storing the imaging timing, the gradient magnetic field is measured by installing an air-core solenoid coil 110 near the gantry of the MRI apparatus 1000. The imaging timing can be known by comparing this data with a sequence chart.

  As another method of storing the imaging timing, an RF wave (RF pulse) trigger may be branched from the circuit of the control unit 230 and stored. On the other hand, the position of the vibrator (speaking organ in this case) is estimated from sound generated by vibration using an optical microphone.

  A large noise is generated from the MRI apparatus 1000 during imaging. If the frequency of the noise and the frequency of the vibrator do not overlap, the sound wave shape including the fundamental frequency of the vibrator can be detected by using the band pass filter 254. . If the sound from the vibrator cannot be measured due to the noise of the scanner, a bone conduction microphone is effective.

If the vibration of the vibrator is not intermittent, it is also necessary to detect the vibration section.
2) The position of the vibrator at the time of RF pulse irradiation is obtained from the data obtained in 1). One period of the vibrator is equally divided in time, and a variable corresponding to an empty k space is prepared for image reconstruction calculation. This corresponds to each frame of the processed moving image. Based on the vibration phase of the vibrator corresponding to each empty k-space, the empty k-space is filled with the raw data lead-out line corresponding to this k-space. In many cases, the vibration frequency is not constant, and the position of the vibrator is not always the same at the time of RF pulse irradiation. Therefore, the lead outline data that minimizes these errors is selected and used.

  3) Image reconstruction is performed on k-space data created for each position (phase) of the transducer, and an MRI image of each frame is obtained. By continuously playing this image, it becomes an MRI movie.

  In FIG. 3, motion artifacts appear in the general method of reconstructing an image by sequentially filling lead outlines (RO-lines) with consecutive n pieces of data, whereas there is no artifact in the method of the MRI apparatus 1000. The shape of the vibrator can be visually recognized.

  FIG. 4 is a diagram showing temporal changes in the cross-sectional shape of the vocal cords, a speech waveform, and a signal waveform after the speech signal has passed through the band-pass filter when the vocal cord is considered as a vocal organ.

  As described above, even when the noise of the MRI apparatus 1000 is present, using the band-pass filter 254 having a predetermined transmission band, the speech signal of the subject 10 can be separated from the voice signal. Is possible.

  One cycle of the speech voice signal of the subject 10 corresponds to one cycle of vocal cord vibration. Therefore, by storing the sound signal from the vibrator (in this case, the vocal cords), the vibration phase of the vibrator can be specified from the vibration phase of the sound signal. In other words, it can be said that the k-space data acquired at the same timing of the vibration phase of the audio signal corresponds to the cross-sectional image of the vibrator at the same vibration phase.

  FIG. 5 is a diagram showing a pulse sequence of the MRI apparatus 1000. FIG. 6 is a flowchart for explaining the processing flow of the MRI apparatus 1000.

Referring to FIGS. 5 and 6, when the process is started, an initial tuning process of the apparatus is performed for each subject 10, and control unit 230 detects the intensity of the electromagnetic field applied from RF coil 104 and reception coil 106. that the signal strength of the adjustment value and the measured center wave numerical f mating inclusive result of ₀ and (calibration value) as the tuning value, is stored in the tuning value storage unit 220 (step S100).

Subsequently, when measurement on the subject 10 is started, the control unit 230 outputs the electromagnetic wave from the RF pulse transmission unit 240 from the RF coil 104 while controlling the gradient magnetic field coil 102, and also selects the slice selection gradient magnetic field G _SS. Is applied (step S102).

Subsequently, the control unit 230 sets the magnitude of the gradient magnetic field for phase encoding (step S104), and controls the gradient coil 102 applies a magnetic field gradient _{G PE} for phase encoding (step S106). On the other hand, the reception processing of the gradient magnetic field variation is executed by the solenoid coil 110 (step S105).

Control unit 230 (step S108) while applying a magnetic field gradient _{G RO} for frequency-encoding (for reading), and a MRI signal from the subject 10, received by the receiving coil 106, amplified by the signal amplifying section 250 ( Step S110), together with the acquisition timing information, is stored in the data storage unit 258.

  On the other hand, the control unit 230 causes the audio signal amplification unit 252 to amplify the signal from the optical microphone 108 and stores the audio signal transmitted through the band-pass filter 254 in the data storage unit 258 together with the acquisition timing information.

  Subsequently, when the process has not been completed a predetermined number of times (step S114), the process returns to step S102. On the other hand, if the processing has been completed a predetermined number of times (step S114), the image reconstruction unit 260 then continues to determine the transducer at that time for each frame of the moving image based on the data acquisition timing information. Data in k space corresponding to the same phase are sorted and combined (step S118).

  The image reconstruction unit 260 reconstructs an image using Fourier transform processing (step S120).

  Subsequently, the image reconstruction unit 260 displays the reconstructed cross-sectional image on the display unit 270 (step S122).

If the measurement has not been completed (step S124), the process returns to step 102 and the measurement is continued. On the other hand, if it is determined that the process has ended, the process ends.
[Demonstration experiment]
(Phantom with vibrator)
In order to confirm the effectiveness of the image reconstruction method of the MRI apparatus 1000 as described above, MR imaging was performed with a phantom with a vibrator that vibrates by air pressure.

FIG. 7 is a diagram for explaining the parts constituting the phantom.
FIG. 8 is a view showing an appearance and a sectional view of the phantom.

  This phantom is composed of two low foam vinyl chloride plates (the right end and the left end in FIG. 7A) and one thermoplastic elastomer (the center in FIG. 7A). On the MRI image, the low foamed vinyl chloride plate has the same brightness value as air, so it appears black and cannot be seen, but the thermoplastic elastomer appears white.

  FIG. 7B shows the dimensions of each part. Each part has a thickness of 5 mm and a substantially square shape with a side of 50 mm. One of the low foamed vinyl chloride plates has a 10 mm square hole in the center, from which air is supplied. Another low foamed vinyl chloride plate is provided with a notch of 30 mm × 20 mm. The appearance after the phantom is assembled is as shown in FIG.

  As shown in FIG. 8B, the thermoplastic elastomer is vibrated by sandwiching the thermoplastic elastomer between the two plates and supplying air from the nozzle.

  In order to measure the position of the vibrator, the optical microphone 108 is used. When the sound was recorded by vibrating the vibrator in a quiet room and the vibration frequency was analyzed, it was confirmed that it was 112.4 Hz. The sound of the vibrator was recorded on the right side of the stereo sound using the optical microphone 108 with a sampling frequency of 44.1 kHz and a quantization number of 16 bits.

  In order to measure the change in the gradient magnetic field of the MRI apparatus 1000, an air-core solenoid coil 110 having a radius of 50 mm was installed at the gantry entrance. The output from this coil was recorded simultaneously on the left side of the stereo sound of the optical microphone 108.

The imaging parameters used are as follows.
1) Sequence: 2D FLASH
2) FOV (imaging area): 128mm x 128mm
3) Matrix: 128 × 128
4) TR: 4.9ms (repetition time)
5) TE: 1.93ms (Echo time)
6) Flip angle: 10 °
7) Bandwidth: 735Hz / pixel
8) Slice thickness: 4mm
9) Number of repetitions: 512
The main component of imaging noise was set to be different from the vibration frequency of the vibrator with the minimum TE.

  As shown in FIG. 8B, the imaging surface is tilted by 20 degrees with respect to the phase direction during MR imaging. This appears in the phase direction of the MRI image when motion artifacts appear on the MRI image, and therefore overlaps the transducer itself in this experiment. In order to avoid this, the angle is set so as not to be parallel to the phase direction.

(Analysis of recorded data and rearrangement of k-space)
The phase of the vibrator at the time of RF pulse irradiation is measured from the recorded voice data. Since the sound recorded by the optical microphone is mixed with noise from the MRI apparatus 1000, the phase of the vibrator at the time of slice excitation cannot be determined directly.

  FIG. 9 is a diagram illustrating a spectrogram of noise generated from the MRI apparatus 1000 and sound generated from the vibrator. From this figure, it can be seen that the fundamental frequency of the phantom vibrator does not overlap with the vibration noise of the scanner. Therefore, the position of the vibrator can be estimated by applying a band-pass filter to the sound generated by the vibrator. As the band-pass filter, a band-pass filter having a cutoff frequency of 100 Hz and 150 Hz, which includes the fundamental frequency of the vibrator and is set so as not to include the noise component of the MRI apparatus 1000, is applied to the position (phase) of the vibrator. Were determined. This filter is once filtered from the front with respect to the acoustic data, and further filtered from the rear. As a result, the phase characteristic of the filter is canceled and the time information of the recorded waveform is retained.

FIG. 10 is a diagram showing the recorded gradient magnetic field and sound.
The upper diagram of FIG. 10 is a gradient magnetic field recorded from an air-core solenoid coil installed beside the gantry of the MRI apparatus 1000, and the lower diagram is a voice waveform recorded by an optical microphone.

  In the upper diagram, a slice selection gradient magnetic field, a phase encoding gradient magnetic field, and a reading gradient magnetic field are detected as the gradient magnetic fields. Furthermore, crusher pulses are also detected.

  In the figure below, the broken line is the original data recorded and contains the acoustic noise of the scanner. The solid line is the waveform after processing by the 100 Hz-150 Hz band-pass filter.

  From the signal of the air-core solenoid coil installed in the gantry, it is possible to confirm the change of the gradient magnetic field. The imaging cycle TR was accurately detected using the first peak in this data as a reference, and these times were used as a reference for rearranging the k-space lead outline.

  Imaging was performed using the method of the MRI apparatus 1000, and an MRI movie was reconstructed. A k-space lead-out was collected from the raw data obtained by scanning. Among the k-space lead-out lines, although not particularly limited, for example, when the fundamental frequency of the phantom vibrates with an error within ± 10% with respect to 112 Hz and ± 20% with respect to the frame center of the moving image An error lead-out of within was used for video reconstruction.

  For verification, we also took a high-speed camera after MRI movie. The frame rate of the high-speed camera was set to 1200 FPS. Images were taken at 336 x 96 pixels, full color 24 bit. In order to match the frame rate of the high-speed camera, the MRI movie was created so that there were 11 frames per period of the transducer.

FIG. 11 is a diagram showing an MRI movie and a high-speed camera image in comparison.
Through such imaging, it was found that the motion of the transducer of the MRI movie well reproduced the actual motion of the transducer.

  In the measurement described above, an MRI movie was obtained in which the maximum number of frames per period of the transducer was 57 frames. In this case, the k-space lead-out was used when the tolerance for the fundamental frequency of 112 Hz of the phantom was ± 10%. In this case, the motion is captured with a time width of 0.16 ms per frame. The frame rate is about 6834 fps (frame per second), and it can be said that the time resolution is very high.

[MRI movie of gumming sounds]
An MRI movie of gum shaking sound using the method of the MRI apparatus 1000 was also created. The test subject is one male. The recording method of MRI imaging parameters, voice, and gradient magnetic field is the same as in the case of phantom. However, the continuous scan was 100 measurements (about 1 minute), and this was repeated 10 times with a 1-minute break. As a result, data for 1000 measurements was obtained.

  The average fundamental frequency of the gum shaking sound was 25.8 Hz. For example, the k-space lead-out outline is within ± 15% error range with respect to this average fundamental frequency. In this way, we were able to obtain an MRI movie with 50 frames per cycle of gum shaking.

  FIG. 12 is a diagram showing data for 26 to 50 frames out of the 50 frames thus obtained. The tongue was in contact with the hard palate between the 33rd and 45th frames. It can be seen that the closing time is 10.1 ms in 38.8 ms of the gum shaking sound cycle, corresponding to 26% of the total.

  As described above, according to the image reconstruction method of the MRI apparatus 1000, an MRI movie with an ultra-high time resolution for imaging a stationary vibrator that vibrates at a high speed such as vocal cords and gum shaking sounds during the utterance period. An imaging method can be realized.

  Therefore, the cross section of the vibrator that vibrates at a high speed about the frequency of sound can be imaged as a moving image.

  The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

  10 subjects, 12 units, 100 static magnetic field coils, 102 gradient magnetic field coils, 104, 106 RF coils, 200 tomography control units, 210 input units, 220 tuning value storage units, 230 control units, 240 RF pulse transmission units, 250 MRI signal amplification unit, 252 audio signal amplification unit, 254 band transmission filter, 258 data storage unit, 260 image reconstruction unit, 270 display unit, 1000 MRI apparatus.

Claims (5)

A magnetic resonance imaging apparatus for detecting a detection signal caused by nuclear magnetic resonance from a transducer to be measured and generating a tomographic motion image of the measurement target,
A static magnetic field applying means for applying a static magnetic field to the measurement object;
Gradient magnetic field applying means for applying a magnetic field, which is modulated so that the detection signal has position information of a nucleus that emits the detection signal, in the selected cross section of the measurement target;
An electromagnetic wave transmitting / receiving means for applying an electromagnetic wave to the measurement object and detecting the detection signal from the measurement object;
Storage means for storing an output signal from the vibrator corresponding to the phase of the periodic motion of the vibrator to be measured in a one-to-one correspondence;
Providing the electromagnetic wave to the electromagnetic wave transmission / reception means, tomographic control means for receiving the detection signal and generating the tomographic moving image as a time-series frame;
The tomography control means includes
Data acquisition means for acquiring data for reconstructing a cross-sectional image multiple times for each row along the lead-out axis of k-space, asynchronously with the motion of the vibrator;
By selecting a row corresponding to the same phase of the motion of the transducer based on the output signal at the timing of each frame of the moving image for reconstructing the cross-sectional image being measured a plurality of times, the phase And an image reconstructing means for reconstructing the image of the magnetic resonance imaging apparatus. Further comprising timing detection means for acquiring information for specifying measurement timing for each row along the readout axis;
The magnetic resonance imaging apparatus according to claim 1, wherein the storage unit stores data for reconstructing the cross-sectional image together with information indicating the measurement timing.   The magnetic resonance imaging apparatus according to claim 1, wherein the output signal is an audio signal from the vibrator.   4. The magnetic resonance imaging apparatus according to claim 1, further comprising a band-pass filter for separating the audio signal from noise from the magnetic resonance imaging apparatus. 5. A magnetic resonance imaging method for detecting a detection signal resulting from nuclear magnetic resonance from a transducer to be measured and generating a tomographic motion image of the measurement target,
Applying a static magnetic field to the object to be measured, and applying a magnetic field, modulated in such a way that the detection signal has position information of the nucleus that emits the detection signal, in a selected cross section of the object to be measured And detecting the detection signal from the measurement object by applying an electromagnetic wave to the measurement object;
Storing an output signal from the transducer corresponding one-to-one with the phase of the periodic motion of the transducer to be measured;
Providing the electromagnetic wave to the object to be measured, receiving the detection signal, and generating the tomographic moving image as a time-series frame,
The step of generating the tomographic image includes
Asynchronously with the motion of the transducer, acquiring data for reconstructing a cross-sectional image multiple times for each row along the k-space readout axis;
By selecting a row corresponding to the same phase of the motion of the transducer based on the output signal at the timing of each frame of the moving image for reconstructing the cross-sectional image being measured a plurality of times, the phase Reconstructing an image of the magnetic resonance imaging method.